Abstrak |
: |
ABSTRAK Informasi curah hujan diperlukan untuk memperkirakan ketinggian air yang masuk ke Daerah Aliran Sungai (DAS). Informasi tersebut dapat diperoleh de- ngan meramalkan curah hujan. Tujuan penelitian ini adalah menentukan model yang digunakan dalam peramalan curah hujan dan klasi_kasi peramalan curah hujan tahun 2015 di DAS Bengawan Solo. Peramalan curah hujan yang mempunyai pola data musiman dapat meng- gunakan model runtun waktu fuzzy dengan pembobot eksponensial. Residu dari model runtun waktu fuzzy terbobot diestimasi menggunakan deret Fourier untuk memperoleh residu yang lebih kecil. Model hibrida runtun waktu fuzzy terbobot- deret Fourier digunakan untuk meramalkan curah hujan tahun 2015. Klasi_kasi peramalan curah hujan di DAS Bengawan Solo ditentukan berdasarkan ukuran Badan Meteorologi Klimatologi dan Geo_sika (BMKG). Berdasarkan hasil penelitian, peramalan curah hujan di DAS Bengawan So- lo dengan model hibrida runtun waktu fuzzy terbobot-deret Fourier menunjukkan Root Mean Square Error (RMSE) yang terkecil. Hasil klasi_kasi peramalan cu- rah hujan tahun 2015 bulan Mei, Juni, Juli, Agustus, September, Oktober dika- tegorikan rendah, bulan November dikategorikan menengah dan bulan Desember dikategorikan tinggi. Kata Kunci : model runtun waktu fuzzy terbobot, deret Fourier, curah hujan, DAS Bengawan Solo ABSTRACT Rainfall information is needed to forecast the water height that ows into river watershed. It's can be obtained by forecasting rainfall. The purpose of this study was to determine the best model among several models in forecasting and classifying 2015's rainfall in Bengawan Solo watershed. The forecasting of rainfall having seasonal data pattern can be using the fuzzy time series model with an exponential weight. Residual of weighted fuzzy time series model was estimated by Fourier series to obtain a smaller residual value. A hybrid model of weighted fuzzy time series-Fourier was used to forecast 2015's rainfall. Classifying rainfall in Bengawan Solo watershed based on Badan Meteorologi Klimatologi dan Geo_sika (BMKG)'s criterion. Based on the result of this research, rainfall forecasting with the hybrid model of weighted fuzzy time series-Fourier showed the smallest Root Mean Squ- are Error (RMSE). Forecast classi_cation showed 2015's rainfall in May, June, July, August, September and October was categorized as low, in November was categorized as average, and in Desember was categorized as high. Keywords : weighted fuzzy time series model, Fourier series, rainfall, Bengawan Solo watershed |